Compiling a fuzz-test with coverage instrumentation

Getting Help
1

I want to build a fuzztest as a standalone binary, for usage in a docker container.

However, the binary complains that the test binary was not built with coverage instrumentation, so fuzzing will run without coverage guidance and may be inefficient.

I want to enable the coverage instrumentation.

Example test:

package main

import (
	"bytes"
	"crypto/sha256"
	"testing"
)

func FuzzFoobar(f *testing.F) {
	f.Fuzz(func(t *testing.T, data []byte) {
		// Invoke some arbitrary methods
		data = sha256.New().Sum(bytes.ToLower(data))
		if len(data) == 0 {
			panic(1)
		}
	})
}

Running it regularly works fine:

$ go test . -fuzz Foo
fuzz: elapsed: 0s, gathering baseline coverage: 0/66 completed
fuzz: elapsed: 0s, gathering baseline coverage: 66/66 completed, now fuzzing with 8 workers
^Cfuzz: elapsed: 3s, execs: 107197 (39642/sec), new interesting: 8 (total: 74)

Compiling it yields complaint:

$ go test  -c .
$ ./testfuzz.test -test.fuzz=Foo -test.fuzzcachedir=/tmp
warning: the test binary was not built with coverage instrumentation, so fuzzing will run without coverage guidance and may be inefficient
warning: starting with empty corpus
fuzz: elapsed: 0s, execs: 0 (0/sec)

If I try to use the -cover when compiling, it does not help either:

$ go test -cover -c .
$ ./testfuzz.test -test.fuzz=Foo -test.fuzzcachedir=/tmp
warning: the test binary was not built with coverage instrumentation, so fuzzing will run without coverage guidance and may be inefficient
warning: starting with empty corpus
fuzz: elapsed: 0s, execs: 0 (0/sec)